Nitric Oxide Decreases Cytokine-induced Endothelial Activation
Nitric Oxide Selectively Reduces Endothelial Expression of Adhesion Molecules
and Proinflammatory Cytokines
Raffaele De Caterina, Peter Libby, Hai-Bing Peng, Victor J. Thannickal, Tripathi B. Rajavashisth,
Michael A. Gimbrone, Jr.,* Wee Soo Shin, and James K. Liao
Vascular Medicine and Atherosclerosis Unit, Cardiovascular Division, Department of Medicine, and *Vascular Research Division,
Department of Pathology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts 02115
Abstract
Introduction
To test the hypothesis that nitric oxide (NO) limits endothelial activation, we treated cytokine-stimulated human saphenous vein endothelial cells with several NO donors and assessed their effects on the inducible expression of vascular
cell adhesion molecule-i (VCAM-1). In a concentrationdependent manner, NO inhibited interleukin (IL) -la-stimulated VCAM-1 expression by 35-55% as determined by
cell surface enzyme immunoassays and flow cytometry. This
inhibition was paralleled by reduced monocyte adhesion to
endothelial monolayers in nonstatic assays, was unaffected
by cGMP analogues, and was quantitatively similar after
stimulation by either IL-la, IL-1f, IL-4, tumor necrosis
factor (TNFa), or bacterial lipopolysaccharide. NO also
decreased the endothelial expression of other leukocyte adhesion molecules (E-selectin and to a lesser extent, intercellular adhesion molecule-i) and secretable cytokines (IL-6
and IL-8). Inhibition of endogenous NO production by LN-monomethyl-arginine also induced the expression of
VCAM-1, but did not augment cytokine-induced VCAM-1
expression. Nuclear run-on assays, transfection studies using various VCAM-1 promoter reporter gene constructs,
and electrophoretic mobility shift assays indicated that NO
represses VCAM-1 gene transcription, in part, by inhibiting
NF-cB. We propose that NO's ability to limit endothelial
activation and inhibit monocyte adhesion may contribute to
some of its antiatherogenic and antiinflammatory properties
within the vessel wall. (J. Clin. Invest. 1995. 96:60-68.) Key
words: nitric oxide * endothelium * adhesion molecules
monocytes * atherogenesis
Nitric oxide (NO)' or closely related molecules account for
the activity of endothelium-derived relaxing factor (1, 2). This
mediator stimulates guanylyl cyclase in smooth muscle cells
leading to vascular relaxation (3). Besides its vasodilatory effects, NO reportedly has many antiatherogenic properties. NO
reduces platelet aggregability (4), limits vascular smooth muscle cell proliferation (5, 6), inhibits leukocyte adhesion to the
endothelium (7, 8), and prevents monocyte chemotaxis (9).
Chronic provision of L-arginine in the diets of hypercholesterolemic rabbits has been reported to improve endothelium-dependent vasodilation and to reduce the extent of atherosclerotic
lesions (10). Local leukocyte recruitment to the vessel wall is
an early step in atherogenesis (11), and endothelial-leukocyte
adhesion molecules such as vascular cell adhesion molecule-i
(VCAM-1), intercellular adhesion molecule-I (ICAM-1), and
endothelial-leukocyte adhesion molecule- I (ELAM-1 or E-selectin) may facilitate recruitment of monocytes and lymphocytes to sites of lesion formation (12-15). Therefore, factors
which modulate the expression of endothelial-leukocyte adhesion molecules and other endothelial-derived cytokines may be
important in modulating lesion formation. This study sought
to determine whether NO can limit endothelial activation by
inhibiting adhesion molecule expression and monocyte adhesion.
This work was presented in abstract form at the 1994 Annual Scientific
Meeting of the American Heart Association, Dallas, TX, on 15 November 1994.
Address correspondence to James K. Liao, Vascular Medicine and
Atherosclerosis Unit, Brigham and Women's Hospital, 221 Longwood
Avenue, LMRC 307, Boston, MA 02115. Phone: 617-732-6628, FAX:
617-732-6961. Raffaele De Caterina's present address is CNR Institute
of Clinical Physiology, Pisa, Italy. Victor J. Thannickal's present address
is Division of Pulmonary and Critical Care Medicine, Tufts-New England Medical Center, Boston, MA. Wee Soo Shin's present address is
The Second Department of Medicine, Tokyo University, Tokyo, Japan.
Receivedfor publication 21 November 1994 and accepted in revised
form 21 March 1995.
J. Clin. Invest.
© The American Society for Clinical Investigation, Inc.
0021-9738/95/07/0060/09 $2.00
Volume 96, July 1995, 60-68
60
De Caterina et al.
Methods
Reagents. Sodium nitroprusside (SNP) was purchased from Elkins-Sinn
(Cherry Hill, NJ), and 3-morpholino sydnonimine (SIN-1) was a gift
from Dr. A. G. Cassella (Frankfurt, Germany). S-nitroso-glutathione
(GSNO) was chemically synthesized from glutathione and sodium nitrite (Sigma Immunochemicals, St. Louis, MO) (16). IL-la was obtained from Hoffmann-La Roche (Nutley, NJ). IL-1,3, TNFa, and IL4 were purchased from Genzyme Corp. (Cambridge, MA). Escherichia
coli lipopolysaccharide (LPS) was purchased from Sigma Immunochemicals. Cell-permeable superoxide dismutase (SOD) and catalase
coupled to polyethylene glycol (PEG) were obtained from Sigma/Enzon
(St. Louis, MO) (17). Assays for IL-6 and IL-8 were measured according to commercially available enzyme immunoassay (EIA) kits
(R & D Systems, Inc., Minneapolis, MN), assayed directly in human
saphenous vein endothelial cell (HSVEC) supernates.
1. Abbreviations used in this paper: CAT, chloramphenicol acetyltransferase; EIA, enzyme immunoassay; GSNO, S-nitroso-glutathione; hpf,
high-power field; HSVEC, human saphenous vein endothelial cells;
ICAM-1, intercellular adhesion molecule-l; L-NAME, L-nitro-arginine
methyl ester; L-NMA, L-N-monomethyl-arginine; NO, nitric oxide;
SIN-1, 3-morpholino sydnonimine; SNP, sodium nitroprusside; VCAM1, vascular cell adhesion molecule-1.
Cell culture. HSVEC were harvested enzymatically with type II
collagenase 0.1% as described (18) and maintained in medium 199
(GIBCO BRL, Gaithersburg, MD), containing Hepes (25 mmol/liter),
heparin (1%), endothelial cell growth factor (50 Ag/m1l), L-glutamine
(1%), antibiotics, and 5% fetal calf serum (Hyclone Laboratories, Logan, UT). Once grown to confluence, the cells were replated on low
pyrogen fibronectin (1.5 tig/cm2) at 20,000 cells/cm2. HSVEC isolated
by these techniques form a confluent monolayer of polygonal cells and
express von Willebrand factor as determined by their content of specific
mRNA and immunoreactive protein (18). Cell number was assessed
after trypsinization in a Neubauer hemocytometer (VWR Scientific
Corp., Philadelphia, PA). Cellular viability was assessed by Trypan
blue exclusion. Bovine aortic endothelial cells (106/100-mm culture
dish, grown in DME with 5% heat-inactivated fetal calf serum) were
used within passage 3 for transfection experiments.
Detection ofcell surface molecules. Assay of cell surface molecules
was carried out after treatment with cytokines and NO donors for 24 h.
The expression of adhesion molecules was determined by cell surface
EIA or flow cytometry using mouse anti-human monoclonal antibodies
against VCAM-1 (antibody El/6), E-selectin (Ab H18/7), ICAM-1
(Ab HU 5/3), MHC-I (W6/32) (19), or a constitutive and non-cytokine-inducible endothelial cell antigen (El/1) (20). These last two
endothelial surface molecules share a high level of basal constitutive
expression, although MHC-I antigen (Ag), at variance from El/I Ag,
shows some degree of inducibility by cytokines. EIA were carried out
by incubating monolayers first with saturating concentrations of specific
monoclonal antibodies against the target molecule, then with biotinylated goat anti-mouse IgG (Vector Labs, Inc., Burlingame, CA), and
finally with streptavidin-alkaline phosphatase (Zymed Laboratories,
Inc., South San Francisco, CA). Layers were washed three times between each incubation step, and integrity of the monolayer was monitored by phase-contrast microscopy. The surface expression of each
protein was quantified spectrophotometrically, reading the optical density of the wells (410 nm) 15-60 min after the addition of a chromogenic
substrate (para-nitrophenylphosphate), as described (19).
Flow cytometric analysis of adhesion molecules was performed by
suspending HSVEC in Hanks' buffered saline solution containing 3
mmol/liter EDTA and incubating the cells with specific primary antibody for 30 min at 4°C. After washing, the cells are then exposed to
goat anti-mouse F(ab')2 labeled with fluorescein-isothiocyanate (Caltag Laboratories, San Francisco, CA) at 4°C. After further washing, the
cell suspension was passed through a FACScan6 analyzer (BectonDickinson Immunocytometry Systems, Mountain View, CA). Results
were plotted as intensity of fluorescence (arbitrary units, on logarithmic
scale, on the abscissa) versus cell number (on the ordinate, total cells
counted, 104 for each condition).
Assessment of total protein synthesis. HSVEC were cultured in 96well plates and stimulated with IL-la in the presence and absence of
NO donors. After 24 h, the total cell-associated protein content was
assessed by the amido black assay, as previously described (21). Briefly,
HSVEC monolayers were fixed with 4% formaldehyde in a 0.1 mol/
liter acetate buffer, pH 3.1, and incubated with amido black B (Sigma
Immunochemicals) to stain cell proteins (22). Unbound dye was rinsed
away with repeated PBS washings, and-dye uptake was determined at
620 nm in an EIA microtiter photometer. In parallel experiments, dye
binding correlated linearly (r = 0.98) with cell number per well, which
in turn was determined by hemocytometric counting of trypsinized cells
in parallel cultures.
Preparation of human monocytes and U937 cells and adhesion
assays. Human peripheral blood monocytes were obtained by centrifugation of white blood cell-enriched blood (Dana Farber Cancer Institute,
Boston, MA) on Ficoll-Hypaque density gradient at 15°C (LMS; Organon Teknika, Durham, NC) followed by counterflow centrifugation
elutriation in a Beckman J2-21 M/E centrifuge using a JE-6 elutriation
rotor and a 6-ml Sanderson chamber (Beckman Instruments, Inc., Palo
Alto, CA). A modified Doherty's method was used (23, 24): the elutriation buffer was Dulbecco's modified phosphate-buffered saline (GIBCO
BRL), supplemented with 3 mmol/liter EDTA and 0.25% human serum
albumin (HSA; Sigma Immunochemicals). The mononuclear cell fraction isolated by Ficoll-Hypaque density gradient at the Ficoll-plasma
interface was loaded at 14 ml/min onto the elutriator centrifuge rotor
head (2,500±10 rpm at 10C), and fractions of elutriated cells were
collected. Monocytes were eluted at 21.5 ml/min. Monocyte suspensions
with this technique are 89±4% pure with 8% lymphocyte, < 2% granulocyte, and essentially no platelet contamination as determined by light
scatter (FACScan0; Becton Dickinson Immunocytometry Systems) and
cell surface antigen analysis with mAb directed to CD14. Monocytes
were resuspended in cold Dulbecco's modified phosphate-buffered saline containing 0.75 mmol/liter Ca2" and 0.75 mmol/liter Mg2" and
0.2% HSA. U937 cells were obtained through American Tissue Culture
Collection (Rockville, MD) and grown in RPMI 1640 (GIBCO BRL)
containing 10% fetal calf serum. Both monocytes and U937 cells were
concentrated by centrifugation at 1 X 106 cells/ml.
For the adhesion assays, HSVEC were grown to confluency in 6well tissue culture plates, after which IL-4 (50 ng/ml) or IL-la (10 ng/
ml) was added for an additional 24 h, in order to induce the expression of
VCAM-1, in the presence or absence of GSNO (0.2 X IO-3 mol/liter).
For control, some monolayers were treated with a mouse anti-human
monoclonal antibody (El /6) against VCAM-1. The adhesion assay was
performed by adding 1 ml of the concentrated monocytes or U937 cell
suspension to each monolayer under rotating conditions (63 rpm) at
21°C. After 10 min, nonadhering cells were removed by gentle washing
with medium 199, and monolayers were fixed with 1% paraformaldehyde. The number of adherent cells was determined by counting six
different high-power fields (hpf) using an ocular grid and a X 20 objective (0.16 mm2/field). Fields for counting adherent leukocytes were
randomly located at half-radius distance from the center of the monolayers.
Northern analysis. HSVEC RNA was isolated by ultracentrifugation
through CsCl (25). Total RNA (20 Isg) was separated on a 1.2%
agarose-formaldehyde gel, transferred to a nylon membrane (Hybond-N;
Amersham Corp., Arlington Heights, IL) and immobilized by shortwave
ultraviolet illumination. cDNA probes were labeled by random priming,
[32P]CTP (3,000 Ci/mmol), and Klenow fragment of DNA polymerase
I (Pharmacia LKB Biotechnology, Inc., Piscataway, NJ). All blots were
hybridized at 42°C overnight and washed (0.2 x SSC, 0.1% SDS, 63°C)
before autoradiography at -80°C for 24-72 h.
Nuclear run-on assay. Nuclei from 3-5 x 107 endothelial cells
were prepared, and in vitro transcription with [a-32P]UTP (800 Ci/
mmol) was performed as described (26). Linearized plasmids (1 4g)
were immobilized on nylon membranes, hybridized to radiolabeled transcripts ( - 5-8 x 107 cpm/ml) at 45°C for 48 h in hybridization buffer
containing 50% formamide, 5 x SSC, 2.5 X Denhardt's Solution, 25
mM sodium phosphate buffer (pH 6.5), 0.1% SDS, and 250 mg/ml
salmon sperm DNA, and washed in 1 x SSC/0.1% SDS at 650C before
autoradiography.
Transfections. The human VCAM-1 promoter constructs containing
the chloramphenicol acetyltransferase (CAT) reporter were described
previously by Neish et al. (27). [ -755 ] FO.CAT is the functional VCAM1 promoter containing AP-1, GATA, and NF-KB binding sites.
[-98]F3.CAT contains the NF-KB binding sites without GATA or AP1. [-44]F4.CAT lacks NF-KB bindings sites. Bovine aortic endothelial
cells were transfected with each reporter plasmid (25 jsg) using the
calcium phosphate precipitation method (28). As an internal control
for transfection efficiency, pRSV./3GAL plasmid (10 /ig) was cotransfected in all experiments. Preliminary results using 3-galactosidase
staining indicate that cellular transfection efficiency was - 12-15%.
Cells (60-70% confluent) were stimulated 48 h after transfection with
IL-la ( 10 ng/ml) and TNFa ( 10 ng/ml) (alone or in combination with
GSNO [0.2 mmol/liter]), and cellular extracts were prepared 12 h later
using lysis buffer (100 mg/ml leupeptin, 50 pg/ml aprotinin, 0.1 mM
PMSF, 5 mM EDTA, 5 mM EGTA, 100 mM NaCl, 5 mM Tris-HCl, pH
7.4) and one freeze-thaw cycle. The cellular extracts were centrifuged at
12,000 g for 10 min, and the supernatant was subjected to CAT and /galactosidase assay.
CAT assays. CAT activity was determined by incubating the superNitric Oxide and Endothelial Leukocyte Adhesion Molecules
61
Table L Lack of GSNO Effects on Various Parameters of HSVEC
Viability and Protein Expression
GSNO
T
0
*
Control
T~~~~~~~~~~
Cell count (cells/103/cm2)
Viability (Trypan blue exclusion)
Total proteins (OD mU)
Constitutive endothelial cell
surface Ag (OD mU)
MHC-1
50
25S
El/I
0
O'
3 x 10,s
1
t4
2 x 10-
1 -3
(200
jumol/liter, 24 h)
56±8
55+6
> 95%
60.6±0.8
> 95%
57.6+2.1
825.5±31.8
1726.5±51.5
1745.0±57.4
823.0±16.7
Values are mean±SEM.
CONCENTRATION (mol/L)
Figure 1. The effect of SNP (black bars), SIN-1 (white bars), and
GSNO (hatched bars) on IL-la ( 10 ng/ml) -stimulated VCAM-1 expression. IL-la and NO donors were coincubated for 24 h before cell
surface EIA. Results are expressed relative to absence of NO donors
(percentage of control). Original readings were derived from four different experiments each consisting of 2 12 different replicates for each
condition. * Significant difference (P < 0.05) compared with controls
and * *significant difference (P < 0.01 ) compared with the effects of
SNP and SIN-1 at 0.4 mol/liter.
natant (100 ul) with [3H]chloramphenicol (50 ,Ci/ml) and n-butyryl
coenzyme A (250 p1g/ml) for 20 h at 37°C (29). The n-butyryl [3H]chloramphenicol was then separated from unmodified chloramphenicol
by xylene phase extraction and counted for 2 min in a liquid scintillation
counter (LS 1800; Beckman Instruments, Inc.). CAT activity was calculated from a standard curve using various concentrations of purified
CAT (Promega, Madison, WI). /1-Galactosidase activity was assayed
spectrophotometrically (absorption at 410 nm) and compared with a
standard curve using known amounts of purified ,/-galactosidase (Sigma
Immunochemicals) as described previously (30). The normalized CAT
activity was calculated as the ratio of CAT to /3-galactosidase activity.
The relative CAT activity was standardized to the normalized FO.CAT
activity having a value of 100. Each experiment was performed three
times in duplicate and all experiments included both positive (highly
expressed pSV40.CAT) and negative (promoterless po.CAT) controls.
Electrophoretic mobility shift assay. Nuclear extracts were prepared
as described (31). The NF-KB oligonucleotide (CCTGGGTTTCCCCTTGAAGGGATTTCCCTCC) spanning the two tandem KB sites in
the human VCAM-1 promoter (27) was synthesized by Genosys Biotechnologies, Inc. (The Woodlands, TX). The oligonucleotide was endlabeled by T4 polynucleotide kinase (New England Biolabs Inc., Beverly, MA) and [y-32P]ATP (3,000 Ci/mmol) and purified by Sephadex
G-50 columns (Pharmacia LKB Biotechnology, Inc.). Nuclear extract
(10 jig) was added to 32P-labeled NF-KB oligonucleotide (- 20,000
cpm) in a buffer containing 2 pg poly [dI * dC] (Boehringer Mannheim
Corp., Indianapolis, IN), 10 tig BSA, 10 mM Tris-HCl (pH 7.5), 50
mM NaCl, 1 mM DTT, 1 mM EDTA, and 5% glycerol. DNA-protein
complexes were resolved on 4% nondenaturing polyacrylamide gel electrophoresed at 12 V/cm in 0.5 x TBE. Specificity was determined
by the addition of p65 or p50 antibodies (15 gg IgG/ml; Santa Cruz
Biotechnology Inc., Santa Cruz, CA) or excess unlabeled (cold) NFKB oligonucleotide (20 ng) to the nuclear extracts for 10 min before
addition of radiolabeled probe.
Statistics. Multiple comparisons were performed by one-way analysis of variance (ANOVA) and individual differences tested by the Fisher's protected least significance difference test after the demonstration
of significant inter-group differences by ANOVA. Student's t test for
62
De Caterina et al.
unpaired data was used to compare the IL-i-stimulated production of
IL-6 and IL-8 with or without GSNO. Comparisons of distribution of
fluorescent intensities at flow-cytometry were performed by the Kolgomorov-Smirnov's statistics (32).
Results
Three structurally unrelated NO donors inhibit IL-la-stimulated VCAM-l expression. SNP, SIN-1, and GSNO all inhibited
IL-la-stimulated expression of VCAM-1 in a dose-dependent
fashion, as assessed by cell surface EIA (Fig. 1). Although
maximal inhibition of adhesion molecule expression required
the presence of the NO donors throughout the period of IL-la
stimulation, coincubation of NO donors for periods as short as
1 h could still inhibit VCAM-1 expression, albeit to a lesser
extent (see below).
Incubation of HSVEC with various NO donors for 24 h
resulted in significantly different toxicity profiles, with maximum subtoxic concentrations (concentrations producing < 10%
decrease in either cell count or total protein synthesis) being
10-4 mol/liter for SNP and SIN-i and 10' mol/liter for
GSNO. At I0-3 mol/liter, SNP and SIN-I significantly reduced
cell number counts (12 and 14%, respectively), while GSNO
did not. GSNO inhibited total protein synthesis (by 9 and 19%)
only at concentrations of 1 x 10-' and 10 x 10' mol/liter,
respectively. Because of its lower toxicity, most subsequent
experiments used GSNO.
NO itself most likely caused the inhibition of VCAM-1
expression by GSNO, since neither of its chemical precursors,
nitrite and glutathione, produced any inhibitory effects at comparable concentrations (data not shown). GSNO (0.2 x 10-'
mol/liter) caused a 35-55% (range) reduction of IL-la-stimulated VCAM-l expression after 24 h, but had no effect on
the surface expression of two constitutive endothelial surface
proteins, El/1 and MHC-I (Table I).
Flow cytometric analysis also demonstrated inhibition of
VCAM-1 expression by GSNO (Fig. 2). Unstimulated HSVEC
expressed little or no VCAM-1, as the distribution of fluorescence intensity compared with cell number in such conditions
(Fig. 2 A) overlapped that obtained with a control preimmune
serum (not shown). Treatment with GSNO did not affect basal
VCAM-1 surface expression (Fig. 2 B). Stimulation of HSVEC
with IL-la shifted the cell population to the right and widened
its distribution (Fig. 2 C), indicating increased VCAM-1 expression in the cell population. Compared with IL-la alone,
1500
CONTROL
3.37 ± 0.42
0
80
10
100
~~~GSNO
=
0
ti4
v:
q
-H
.
* control
0 GSNO
1000-
g*
Ed 500-
CELL
NUMBER
0
100
10
80
IL-1
C
IC
80 1 o
l|
Do
i A4
10
,
~~~~~~~~IL-I
~~~13.6±i1.6
IL-1 + GSNO
10.0 ± 1.4
100
RELATIVE FLUORESCENCE INTENSflY
Figure 2. Flow cytometric analysis of endothelial cells using antibody
to VCAM-1. (A) Control (no IL-la, no GSNO); (B) the effect of
GSNO (0.2 x 10-3 mol/liter) after 24 h; (C) the effect of IL-la (10
ng/ml) after 24 h; and (D) combination of IL-la and GSNO. Plots
represent the distributions of relative fluorescence intensities (in arbitrary units, or "channels," on the abscissas, on a logarithmic scale) as
a function of cell number (on the ordinates). Cell number analyzed in
each condition was 10,000. Insets represent the median values of the
distribution±SEM. Panels shown are from one of two separate experiments using different sets of HSVEC monolayers.
GSNO, when added with 11L-la, shifted the distribution leftward, decreasing both its median value and its dispersion (Fig.
2 D). These results indicate that, under GSNO treatment, fewer
cells expressed VCAM-1 and, on the average, each cell did so
to a lesser extent. Controls for flow cytometry included the
use of an antibody against a constitutive and non-cytokineinducible endothelial antigen (Ab El /1), with which all distributions appeared uniformly shifted to the right, and a preimmune serum (results not shown), which yielded distributions
similar to those shown in Fig. 2, A and B.
GSNO inhibits other cytokine-stimulated cell-associated
and secreted molecules. We investigated the effect of GSNO
on other cytokine-inducible endothelial proteins, including two
other endothelial-leukocyte adhesion molecules, E-selectin and
ICAM-1, and two secretable proteins, IL-6 and 11L-8. GSNO
inhibited cytokine-induced surface expression or secretion of
each of these proteins (Fig. 3 and Table II). In comparative
experiments, inhibition of inducible ICAM-1 expression, however, was negligible compared with that of either VCAM-1 or
E-selectin (Fig. 3). GSNO did not affect basal (constitutive)
ICAM-1 expression. Inhibition of IL-8 also was lower than for
IL-6 (Table II). These experiments indicate that NO selectively
decreases production of certain cytokine-induced molecules in
endothelial cells.
GSNO inhibits VCAM-J expression in response to other
cytokines. We assessed and compared the inhibition of VCAM1 expression by GSNO in response to a variety of structurally
unrelated agonists such as IL-1p3, IL-4, TNFa, and LPS. GSNO
inhibited VCAM-1 expression to virtually the same extent with
each stimulus tested (Fig. 4), independent of the relative potency of the stimulus.
Maximum inhibition of VCAM-J by GSNO requires pro-
0~ O-
_
+
VCAM-1
-
+
-
E-selectin
+
T
I
MA
ICAM-1
Figure 3. Comparative expression of VCAM-1, E-selectin, and ICAM1 in HSVEC in the presence or absence of IL-la (10 ng/ml, 24 h),
with or without GSNO (0.2 x 10'- mol/liter), as assessed by cell
surface ETA. Data represent 2 12 replicates for each condition. * Statistically significant differences (P < 0.05) between groups joined by the
horizontal lines.
longed incubation. GSNO inhibited IL-la-stimulated VCAM1 expression after 1 h. However, the continuous presence of
this NO donor during cytokine stimulation produced greater
inhibition. In experiments in which exposure to the NO donor
was stopped at various time points after the coaddition of the
NO donor and IL-la, GSNO inhibited VCAM-1 expression by
14, 27, and 45% with 1, 6, and 24 h of coincubation, respectively. The effect of SNP and SIN-I at the same concentrations,
on the other hand, did not increase beyond the level observed
after 1 h (38% inhibition).
GSNO decreases monocyte adhesion. Monocytes adhered
little to unstimulated HSVEC (control) under nonstatic conditions at 21°( (Fig. 5, left). Exposure of HSVEC to IL-la (10
ng/ml, 24 h) increased monocyte adhesion substantially (Fig.
5, middle). Treatment of HSVEC with GSNO (0.2 X 10-3
mol/liter) inhibited IL-la-induced monocyte adhesion by 61%
(Fig. 5, right). 11-4 yielded qualitatively similar results (baseline: 2±1 cells/hpf; IL-4 50 ng/ml: 27±5 cells/hpf; 11L-4
+ GSNO: 8±3 cells/hpf, corresponding to a 70% inhibition).
Similar effects were also observed for U937 cell adhesion (not
Table II. The Effect of NO (GSNO, 200
on the Secretion of IL-6 and IL-8
pmolliter, 24 h)
IL-8
IL-6
Inhibition
pg/105 cells
< 20
Control, no IL-la
< 20
Control + GSNO
IL-la (10 ng/ml) 286.0±19.8
IL-la + GSNO
142.0±3.3*
by GSNO
ND
50%
pg/105 cells
25.9±3.0
25.8±1.3
970.5±31.0
723.1±46.7*
Inhibition
by GSNO
< 1%
25%
Values are mean±SEM of protein concentrations assayed in quadruplicate in medium from greater than four HSVEC monolayers in each
control unstimulated condition and greater than eight HSVEC monolayers for each stimulated condition. Experimental conditions are similar
to what is described for assay of surface-associated molecules. ND, not
detectable. * P < 0.01 in comparison with corresponding control conditions.
Nitric Oxide and Endothelial Leukocyte Adhesion Molecules
63
C**
O
W3+1
p
*
**
1000
800
Cj
600
m
400
* control
0 GSNO 200 jwmol/L
200
no stim. IL-la IL-11
TNFa IL-4
LPS
shown). These observations parallel the decrease in leukocyte
adhesion molecule expression caused by GSNO.
Inhibition of VCAM-J expression by GSNO is not mediated
by cGMP. To test whether cGMP mediates the inhibitory effect
of NO donors on VCAM-1 levels, we tested the effects on ILla-stimulated VCAM-l expression of two membrane-permeable cGMP analogues, dibutyryl cGMP and 8-bromo cGMP, in
a range of concentrations from 10-8 to IO-5 mol/liter. We chose
these concentrations because they did not produce toxic effects
on endothelial cells (as judged by morphology, Trypan blue
exclusion, and cell counting) during the 24-h incubation and
because they encompass concentrations able to mimic vasodilatory and antiplatelet effects of NO (3). Neither agent reduced
IL-la -stimulated VCAM expression (Fig. 6 A). Furthermore,
there was no trend toward any decline over the entire range of
concentrations explored. Preliminary studies using metabolic
labeling with [ y-32P]ATP demonstrated that 8-bromo-cGMP
(1 X 10-3 mol/liter) was effective in activating cGMP-dependent protein kinase activity in endothelial cells, despite lacking
effects on cytokine-induced VCAM-1 expression.
NO decreases VCAM-J mRNA levels. Northern analysis
demonstrated decreased steady state VCAM-1 mRNA levels
upon incubation with either SNP or GSNO at different time
CONTROL
2 +1
IL-la
427 ± 62
Figure 4. The inhibition by GSNO (0.2
X 10-3 mol/liter for 24 h) of VCAM-1 expression by HSVEC in response to different
stimuli: IL-la (10 ng/ml), IL-1,i (10 ng/
ml), TNFa (10 ng/ml), IL-4 (50 ng/ml),
and LPS (1 pg/ml), all administered for 24
h. n = 16 for each different condition. * Statistically significant differences (P < 0.05)
between groups joined by the horizontal
lines.
points after IL-la addition (Fig. 6 A). Densitometric analysis
of autoradiographic bands showed a 56-70% decrease compared with control levels, in good agreement with the fall in
protein expression as quantified by cell surface EIA. Furthermore, GSNO (0.2 X i0' mol/liter) diminished VCAM-1
mRNA expression over a range of IL-la concentrations (0.1100 ng/ml) (Fig. 6 B). Inhibition of endogenous NO with LN-monomethyl-arginine (L-NMA; 1 X i0' mol/liter) augmented the expression of VCAM-1 mRNA in unstimulated
cells, but had lesser effect on VCAM-1 mRNA expression in
TNFa-stimulated cells (Fig. 6 C). In addition, the induction of
VCAM-1 mRNA expression by TNFa was blunted by cellpermeable PEG-catalase, and to a lesser extent, PEG-SOD, suggesting involvement of hydrogen peroxide to a greater extent
than superoxide anion in TN'Fa-induced modulation of VCAM1 mRNA (Fig. 6 D).
NO decreases cytokine-stimulated VCAM-1 transcription.
In the presence of actinomycin D, GSNO did not significantly
affect the half-life of VCAM-1 mRNA (5.2±+1.8 h vs. 4.7±1.5
h, P = NS) indicating that the inhibitory effect of NO on
VCAM-1 expression occurs at the level of transcription (Fig.
7 A). Nuclear run-on assays demonstrated a transcriptional effect of NO on VCAM-1 expression (Fig. 7 B). Transcription
IL-l a + GSNO
165±47
Figure 5. The inhibition by GSNO (0.2 x 10-3 mol/liter for 24 h) of monocyte adhesion to HSVEC induced by IL-la, coadministered with GSNO
for 24 h. Photographs represent randomly chosen fields at half-radius distance from the center of the well in one of three similar comparative
experiments. Insets represent monocyte cell counts±SEM (n = 6 for each condition) within a high-power field (magnification grid area of 0.16
mm2)
64
De Caterina et al.
-* + S
SNP
GSNO
++SNO
A
Relative Intensity
++
0.3 1.0
6.3152
0.3 1.0
;-
7.115.9
_
If
t
Figure 6. (A) Northern analyses
(20 jg total RNA/lane) showing time-dependent effects of
(1lo- molffiter) and
~~~~~~~SNP
SNP (lo-4mol/liter)and
GSNO (l0'~mol/liter) on ]ILleta(lO ng/ml)-induced
2ss
VCAM-1 mRNA levels. Band
intensities were analyzed by
_GSNO
_
densitometry and expressed rell_
t
28 S
1.
ative to the band intensity corre6
1
6
time after IL-la (h)
1
IL-la (10 ngmi) 0 0.1 1.0 10 100 sponding to IL-la-induced ex-
VCAM-1
pression of VCAM-l
D
each set of experiments (relative
intensity). The lower panel
shows the corresponding ethidium bromide staining for 28 S
ribosomal RNA. Three separate
experiments yielded similar results. (B) Northern analyses (20
28S-
Control
C
L-NMA
at 1 h for
-g total RNA/lane) showing
-.t28S
effects
:concentration-dependent
_
SOD
of IL-la on VCAM-l mRNA
-.;f--^ Em Vexpression in the presence and
les -
arm-288-X|
TN~a
absence of GSNO (0.2 mM).
g
Another
independent experiment yielded similar results. (C)
Northern analyses showing
time-dependent effects of LNMA (1 mM, 20 Mg total RNA/
TNFa + L-NMA
lane) or TNFa (10 ng/ml, 10
o 0.5 2 6 12
0 0.5 2 6 12
MgtotalRNA/lane)alone, orin
Time (h)
Time (h)
combination (10 Mg total RNA/
lane), on VCAM-l mRNA expression. Another independent experiment yielded similar results. (D) Northern analyses (20 tg total RNA/lane) showing time-dependent effects
of TNFa (Control) on VCAM-l mRNA expression in the presence of PEG-SOD (3,000 U/ml) or PEG-catalase (40,000 U/ml). Another separate
experiment yielded similar results.
of the 3-tubulin gene served as an internal control, and lack
of hybridization to the insertless vector, pGEM, established
specificity (data not shown).
To localize possible sites of inhibition on the VCAM-1
promoter by NO, we performed transfection studies with various
deletional promoter reporter CAT constructs (Fig. 8 A) using
bovine aortic endothelial cells which yielded a higher transfection efficiency than human endothelial cells. The promoterless
po.CAT and the highly expressed pSV40.CAT served as negative and positive controls in the reporter assays. The promoterless po.CAT and pF4.CAT produced essentially no relative
CAT activity either basally or after stimulation with TNFa (Fig.
8 B). Stimulation with TNFa of transfectants containing the
fully functional VCAM-1 promoter, FO, increased relative CAT
activity from 101±16 to 1,240±235. Cotreatment with GSNO
resulted in a 76% decrease in relative CAT activity (300±63),
indicating that NO can inhibit TNFa-stimulated VCAM-1 promoter activity.
Transfection studies were also performed using pF3.CAT,
a deletional VCAM-I promoter construct containing two tandem NF-KB sites, but no AP-1 and GATA sites (Fig. 8 B).
Stimulation of pF3.CAT transfectants with IL-la and TNFa
increased relative CAT activity from 73±23 to 474±52, suggesting that binding of NF-KB may be involved in the induction
of VCAM-1 transcriptional activity, albeit at a significantly
lower level compared with pFO.CAT. Cotreatment with GSNO
reduced the relative CAT activity of pF3.CAT transfectants to
basal levels (139±51).
NO inhibits NF-KB activation. To determine whether NO
inhibits NF-KB activation, electrophoretic mobility shift assays
were performed using radiolabeled oligonucleotide corresponding to the two tandem KB sites in the VCAM-1 promoter (Fig.
9). GSNO (0.2 x 10-3 mol/liter) decreased the amount of
shifted complex induced by TNFa, indicating that NO inhibits
the activation of NF-KB. The shifted complexes were specific
for NF-KB since they were supershifted in the presence of antibodies to NF-KB subunits and disappeared with excess unlabeled oligonucleotide. In addition, neither 8-bromo-cGMP (1
X 10-3 mol/liter), glutathione (0.2 x 10-3 mol/liter), nor nitrite (0.2 x 10-3 mol/liter) had any effects on activation of
NF-KB (data not shown).
Discussion
We have shown that NO can reduce cytokine-induced expression of a number of pathophysiologically relevant effector molecules characteristic of endothelial activation. Upon appropriate
stimulation, the endothelium exhibits increased adhesiveness
for monocytes, lymphocytes, and granulocytes. Increased or
newly induced expression of endothelial-leukocyte adhesion
molecules (e.g., ICAM-1, VCAM-1, and E-selectin) mediates
such enhanced adhesive interactions ( 13, 14, 33, 34). The secreNitric Oxide and Endothelial Leukocyte Adhesion Molecules
65
A
A
FO
(-75s)
AP-1
GATA
F3
TATAA
NF-xB
(.95)
NF- 44B TA)
A
F4
TATAA
B
TIME (h)
_CA_
B
_
_IL-la
-
P-Tubulln
CONTROL TNFd TNFz+ NO
l
Figure 7. (A) Half-life of
ulated
CONTROL
( 10 ng/ml) -stim-
12 TNFax
VCAM-1
El
mRNA
and ab-
in the
presence
sence
of GSNO (0.2
mol/liter). Actinomycin D (5 ttg/ml) and GSNO were added 3.5 and 4 h after IL-la
stimulation, respectively. The amount of VCAM-l mRNA (by densitometry) was compared with the amount of VCAM-l mRNA at 4 h
after IL-la stimulation (relative intensity) and plotted as a logarithmic
function of time (h). Time 0 corresponds to 4 h after IL-la stimulation.
(B) Nuclear run-on assay showing the effects of NO (GSNO, 0.2 x 10-3
mol/liter) on TNFa-stimulated VCAM-l gene transcription at 4 h. Band
intensities were normalized to that of /3-tubulin. There were no bands
observed with linearized pGEM plasmid (data not shown). This is
representative of two separate experiments.
TNFhx+GSNO
X 10-3
tion of inflammatory cytokines (IL-6) (34) and of leukocyte-
specific chemoattractants, such as IL-8 and MCP-1, also likely
contributes to leukocyte recruitment during inflammatory responses in vivo (23, 35). NO significantly inhibits the expression of VCAM-1, E-selectin, and IL-6, and, to a lesser extent,
IL-8 and ICAM-1. Three structurally different NO donors inhibited cytokine-induced VCAM-1 expression and concomitantly
reduced monocyte adhesion to cultured endothelial monolayers.
The lack of effects of nitrite or glutathione renders it unlikely
that our findings were due to mediators other than NO.
The physiologic relevance of NO's inhibitory effects is supported by the finding that inhibition of endogenous NO by LNMA can also lead to the induction of VCAM-1 expression,
albeit to a lesser degree compared with cytokine activation.
This result suggests that endogenous endothelial NO production
tonically inhibits VCAM-1 expression in endothelial cell cultures. Because induction with cytokines results in much higher
levels of VCAM-1 expression compared with L-NMA, L-NMA
produced only a small augmentation of VCAM-1 expression in
cytokine-induced endothelial cells. These results agree with our
finding that endogenous levels of NO are not sufficient in limiting cytokine-induced VCAM-1 expression. However, at sites
of inflammation, several cell types such as macrophages and
vascular smooth muscle cells, have the capacity to generate
100-fold higher concentrations of inducible NO at levels comparable with the amount of NO released by GSNO in our study
(36, 37). Furthermore, higher concentrations of NO could be
66
De Caterina et al.
po.CAT
pSV4O
CAT
pFO.CAT
pF3.CAT
pF4.CAT
Figure 8. (A) VCAM- 1 promoter CAT reporter constructs used in transfection experiments. The length of the promoter upstream from the
initiation start site is denoted in base pairs along with putative nuclear
protein binding domains. (B) Transfection studies in bovine aortic endothelial cells using promoterless construct (po.CAT), highly expressed
construct (pSV40.CAT), and various truncated VCAM-1 promoter constructs. * Significant difference (P < 0.01) between treatment with
cytokines alone and in combination with GSNO.
encountered by cytokine-activated endothelial cells due to their
close proximity to smooth muscle cells and macrophages at
sites of inflammation and the potential stabilization of NO by
formation of potent adducts (38).
The mechanism by which NO inhibits cytokine-induced expression of VCAM-1 does not appear to involve stimulation of
guanylyl cyclase since treatment with cGMP analogues does
not reduce VCAM-1 expression. Nor does NO affect the posttranscriptional half-life of VCAM-1 mRNA, as demonstrated
by the actinomycin D studies. Rather, nuclear run-on assays
and transfection studies demonstrate that NO represses VCAM1 gene transcription, in part, by inhibiting nuclear binding protein NF-KB. However, the finding that [-755] FO.CAT had a
higher CAT activity in response to TNFa compared with
[-98]F3.CAT which contained NF-KB sites, but not AP-1 or
GATA sites, suggests that full transcriptional activity of the
VCAM-1 promoter requires the participation of nuclear binding
proteins such as c-fos, c-jun, and GATA.
Inhibition of [-98]F3.CAT and NF-KB by NO suggests that
inactivation of NF-KB may be one of the mechanisms underlying NO's inhibitory effects on VCAM-1 expression. Indeed,
other investigators have found that activation of NF-KB plays
30 min
2h
|~~~~~
r
0-.-
Supershift
-
-NF-cB
41
~
~
Figure 9. Electrophoretic mobility shift assay showing
GSNO
by
tide
mol/liter)
xiO
on
(10 ng/ml). Specificity
TNFa
p50
or
(0.2
(supershift)
antibodies
(20 ng)
or
time-dependent
was
determined
unlabeled
to the nuclear extracts. Three
the effects of
activation of NE-KB
by
addition of
p65
(cold) NE-KB oligonucleo-
separate experiments yielded
similar results.
a
major
transcriptionally inducing
role in
hesion molecules and various
I1L-6
is
IIL-8 (39, 40).
and
thought
to occur, in
the expression of adinflammatory cytokines such as
Since the activation of NF-KB
by
TNFa
part, via reactive oxygen species (41,
42), NO may inhibit NF-KB by scavenging and inactivating
superoxide
one
anion. The reaction of NO and
of the fastest reactions known in
superoxide
and SOD
(43). However,
superoxide
addition of SOD may not
hydrogen peroxide.
Our
extent, SOD, inhibited
at
biological pH,
anion may be
appreciably
findings
superoxide
and is
anion
the spontaneous
sufficiently rapid
that
alter the formation of
that catalase, and to
cytokine-induced
a
lesser
VCAM-1 mRNA
ex-
pression agrees with reports that NF-KB is a more hydrogen
peroxide -sensitive rather than superoxide anion -sensitive transcription
The
of NO to inhibit monocyte adhesion to endothe-
co-workers (9)
endothelium
CDL18,
We are grateful to Maria Muszynski (cell culture), Peter Lopez (flow
cytometry studies), Francis W. Luscinskas (monocyte preparations),
Tucker Collins and Andrew Neish (VCAM-1 promoter constructs),
Peter LoMedico (for providing IL-la), and David G. Harrison (for
helpful discussions regarding GSNO).
This work was supported by National Institutes of Health grants
HL-48743 (P. Libby), HL-36028 (M. A. Gimbrone, Jr.), and HL-05280
(J. K. Liao), by the American Heart Association Grant-in-Aid Award
(J. K. Liao), and by contributions from Italian Consiglio Nazionale
delle Ricerche and Fondazione per la Ricerca Medica (R. De Caterina).
factor (44).
ability
lium. is consistent with previous observations (7, 8). Bath and
to
Acknowledgments
anion is
biological systems
almost three times faster than that between
dismutation of
and collaborators (7) also showed that such inhibitors increase
leukocyte (mostly neutrophil) adhesion to mesenteric venules.
They also showed that increased expression of CDllb/CD18
on the leukocyte surface could not account for L-NAME's effect
(7). Other in vivo studies by Kurose and colleagues (45) have
also documented modulation of monocyte adhesion to endothelial cells by L-NAME and reversal by cGMP analogues. This
result contrasts with our finding that cGMP analogues do not
affect cytokine-induced expression of VCAM-1. Several possibilities could account for this difference. First, these in vivo
studies showed induction of endothelial-leukocyte adhesion by
L-NAME rather than by cytokines. Perhaps induction of
VCAM-1 by L-NAME is amenable to reversal by cGMP analogues. Second, these studies cannot separate cGMP effects
on hemodynamics and/or leukocytes from direct alterations in
endothelial functions. Furthermore, induction of other adhesion
molecules (i.e., ICAM-1 or E-selectin) by L-NAME treatment
may depend partially on cGMP-dependent pathways. In the
present study, by selectively exposing cultured endothelial cells
to NO, we have identified endothelial cells as a potential target
of NO's effect in limiting leukocyte adhesion to VCAM-1.
As mentioned previously, several lines of evidence, both in
vitro and in vivo, have recently suggested a role for endogenous
NO as an antiatherogenic autacoid (46). Our in vitro study
provides a link between NO production and monocyte adhesion
and offers clues as to the mechanism of such an effect. Local
generation of sufficiently high amounts of NO by either endothelial cells themselves or by other neighboring cells in response
to inflammatory stimuli could provide a homeostatic regulation
of monocyte adhesion under physiological conditions within the
vessel wall. Endothelial dysfunction resulting in reduced level
of NO activity can lead not only to an imbalance in vascular
tone, favoring acute vasoconstriction ( 1, 2), but can also impair
an endogenous negative feedback loop that limits VCAM-1
expression, leukocyte adhesion, and atherogenesis.
In summary, we have shown that exogenous as well as
endogenous sources of NO can limit the degree of endothelial
activation. The ability of NO to inhibit the expression of endothelial-leukocyte adhesion molecules and certain proinflammatory cytokines makes it a potentially important regulator of
inflammatory trafficking within the vessel wall.
one
leukocyte
reported
that NO inhibits monocyte adhesion
in vitro, without
of the cognate
altering expression
ligands
adhesion molecules. In
on
of
CDlI lb/
monocytes for endothelial-
experiments superfusing
a cat
mesenteric
preparation with inhibitors of NO production (L-
NMA and
L-nitro-arginine methyl
ester
[L-NAME]), Kubes
References
1. Ignarro, L. J. 1990. Biosynthesis and metabolism of endothelium-derived
nitric oxide. Annu Rev. Pharmacol. Toxicol. 30:535-560.
2. Moncada, S., R. M. J. Palmer, and E. A. Higgs. 1991. Nitric oxide: physiology, pathophysiology, and pharmacology. Pharmacol. Rev. 43:109-142.
3. Ignarro, L. J. 1991. Heme-dependent activation of guanylate cyclase by
nitric oxide: a novel signal transduction mechanism. Blood Vessels. 28:67-73.
4. Radomski, M. W., R. M. Palmer, and S. Moncada. 1990. An L-arginine/
Nitric Oxide and Endothelial Leukocyte Adhesion Molecules
67
nitric oxide pathway present in human platelets regulates aggregation. Proc. Natl.
Acad. Sci. USA. 87:5193-5197.
5. Garg, U. C., and A. Hasid. 1989. Nitric oxide-generating vasodilators and
8-bromo-cycic guanosine monophosphate inhibit mitogenesis and proliferation
of cultured rat vascular smooth muscle cells. J. Clin. Invest. 83:1774-1777.
6. Nakaki, T., M. Nakayama, and R. Kato. 1990. Inhibition by nitric oxide
and nitric oxide-producing vasodilators of DNA synthesis in vascular smooth
muscle cells. Eur. J. Pharmacol. 189:347-353.
7. Kubes, P., M. Suzuki, and D. N. Granger. 1991. Nitric oxide: an endogenous
modulator of leukocyte adhesion. Proc. Nati. Acad. Sci. USA. 88:4651-4655.
8. Tsao, P. S., L. M. McEvoy, H. Drexler, E. C. Butcher, and J. P. Cooke.
1994. Enhanced endothelial adhesiveness in hypercholesterolemia is attenuated
by L-arginine. Circulation. 89:2176-2182.
9. Bath, P. M. W., D. G. Hassal, A.-M. Gladwin, R. M. J. Palmer, and J. F.
Martin. 1991. Nitric oxide and prostacycin: divergence of inhibitory effects on
monocyte chemotaxis and adhesion to endothelium in vitro. Arterioscler. Thromb.
11:254-260.
10. Cooke, J. P., A. H. Singer, P. Tsao, P. Zera, R. A. Rowan, and M. E.
Billingham. 1992. Antiatherogenic effects of L-arginine in the hypercholesterolemic rabbit. J. Clin. Invest. 90:1168-1172.
11. Li, H., M. I. Cybulsky, M. A. Gimbrone, Jr., and P. Libby. 1993. An
atherogenic diet rapidly induces VCAM-1, a cytokine-regulatable mononuclear
leukocyte adhesion molecule, in rabbit aortic endothelium. Arterioscler. Thromb.
13:197-204.
12. Bevilacqua, M. P., J. S. Pober, D. L. Mendrick, R. S. Cotran, and M. A.
Gimbrone, Jr. 1987. Identification of an inducible endothelial-leukocyte adhesion
molecule. Proc. Natl. Acad. Sci. USA. 84:9238-9242.
13. Springer, T. A. 1990. Adhesion receptors of the immune system. Nature
(Lond.). 346:425-434.
14. Cybulsky, M. I., and M. A. Gimbrone, Jr. 1991. Endothelial expression
of a mononuclear leukocyte adhesion molecule during atherogenesis. Science
(Wash. DC). 251:788-791.
15. Springer, T. A. 1994. Traffic signals for lymphocyte recirculation and
leukocyte emigration: the multistep paradigm. Cell. 76:301-314.
16. Kowaluk, E. A., and H. L. Fung. 1990. Spontaneous liberation of nitric
oxide cannot account for in vitro vascular relaxation by S-nitrosothiols. J. Pharmacol. Exp. Ther. 255:1256-1264.
17. Walther, F. J., A. B. Wade, D. Warburton, and H. J. Forman. 1991.
Augmentation of superoxide dismutase and catalase activity in alveolar type II
cells. Am. J. Respir. Cell Mol. Biol. 4:364-368.
18. Libby, P., J. M. Ordovas, K. R. Auger, A. H. Robbins, L. K. Birinyi, and
C. A. Dinarello. 1986. Endotoxin and tumor necrosis factor induce interleukin-1
gene expression in adult human vascular endothelial cells. Am J. Pathol. 124:179185.
19. Warner, S. J. C., G. B. Friedman, and P. Libby. 1989. Regulation of
major histocompatibility gene expression in human vascular smooth muscle cells.
Arteriosclerosis. 9:279-288.
20. Pober, J. S., M. A. Gimbrone, Jr., L. A. Lapierre, D. L. Mendrick, W.
Fiers, R. Rothlein, and T. A. Springer. 1986. Overlapping patterns of activation
of human endothelial cells by interleukin 1, tumor necrosis factor, and immune
interferon. J. Immunol. 137:1893-1896.
21. Libby, P., S. J. C. Warner, and G. B. Friedman. 1988. Interleukin 1: a
mitogen for human vascular smooth muscle cells that induces the release of
growth-inhibitory prostanoids. J. Clin. Invest. 81:487-498.
22. Kohase, M., D. Henriksen-De Stefano, L. T. May, J. Vilcek, and P. B.
Sehgal. 1986. Induction of /32-interferon by tumor necrosis factor: a homeostatic
mechanism in the control of cell proliferation. Cell. 45:659-666.
23. Doherty, D. E., G. P. Downey, G. S. Worthen, C. Haslett, and P. M.
Henson. 1988. Monocyte retention and migration in pulmonary inflammation;
requirement for neutrophils. Lab. Invest. 59:200-213.
24. Luscinskas, F. W., G. S. Kansas, H. Ding, P. Pizcueta, B. E. Schleiffenbaum, T. F. Tedder, and M. A. Gimbrone, Jr. 1994. Monocyte rolling, arrest and
spreading on IL-4-activated vascular endothelium under flow is mediated via
sequential action of L-selectin, fil-integrins, and /32-integrins. J. Cell Biol.
125:1417-1427.
68
De Caterina et al.
25. Chirgwin, J. M., A. E. Przybyla, R. J. MacDonald, and W. J. Rutter.
1979. Isolation of biologically active ribonucleic acid from sources enriched in
ribonuclease. Biochemistry. 18:5294-5299.
26. Liao, J. K., W. S. Shin, W. Y. Lee, and S. L. Clark. 1995. Oxidized lowdensity lipoprotein decreases the expression of endothelial nitric oxide synthase.
J. Biol. Chem. 270:319-324.
27. Neish, A. S., A. J. Willliams, H. J. Palmer, M. Z. Whitley, and T. Collins.
1992. Functional analysis of the human vascular cell adhesion molecule 1 promoter. J. Exp. Med. 176:1583-1593.
28. Graham, F. L., and A. J. Van der Eb. 1973. Transformation of rat cells
by DNA of human adenovirus 5. Virology. 52:456-457.
29. Seed, B., and J.-Y. Sheen. 1988. A simple phase-extraction assay for
chloramphenicol acyltransferase activity. Gene (Amst.). 67:271-277.
30. Edlund, T., M. D. Walker, P. J. Barr, and W. J. Rutter. 1985. Cell-specific
expression of the rat insulin gene: evidence for role of two distinct 5' flanking
elements. Science (Wash. DC). 230:912-916.
31. Dignam, J. D., R. M. Lebovitz, and R. G. Roeder. 1983. Accurate transcription initiation by RNA polymerase H in a soluble extract from isolated mammalian
nuclei. Nucleic Acids Res. 11:1475-1489.
32. Zar, J. 1974. Biostatistical Analysis. Prentice-Hall, Inc., Englewood Cliffs,
NJ. 1-736.
33. Luscinskas, F. W., M. I. Cybulsky, J.-M. Kiely, C. S. Peckins, V. M.
Davis, and M. A. Gimbrone, Jr. 1991. Cytokine-activated human endothelial
monolayers support enhanced neutrophil transmigration via a mechanism involving both endothelial-leukocyte adhesion molecule-i (ELAM-1) and intercellular
adhesion molecule-i (ICAM-1). J. Immunol. 146:1617-1625.
34. Van Snick, J. 1990. Interleukin-6: an overview. Annu. Rev. Immunol.
8:253-278.
35. Oppenheim, J. J., C. 0. Zachariae, N. Mukaida, and K. Matsushima. 1991.
Properties of the novel proinflammatory supergene "intercrine" cytokine family.
Annu. Rev. Immunol. 9:617-648.
36. Kowaluk, E. A., and H.-L. Fung. 1990. Spontaneous liberation of nitric
oxide cannot account for in vitro vascular relaxation by S-nitrosothiols. J. Pharmacol. Exp. Ther. 255:1256-1264.
37. Radomski, M. W., D. D. Rees, A. Dutra, and S. Moncada. 1992. S-nitrosoglutathione inhibits platelet activation in vitro and in vivo. Br. J. Pharmacol.
107:745-749.
38. Stamler, J. S., D. I. Simon, J. A. Osborne, M. E. Mullins, 0. Jaraki, T.
Michel, D. J. Singel, and J. Loscalzo. 1992. S-nitrosylation of proteins with nitric
oxide: synthesis and characterization of biologically active compounds. Proc.
Natl. Acad. Sci. USA. 89:444-448.
39. Libermann, T. A., and D. Baltimore. 1990. Activation of interleukin-6
gene expression through the NF-kappa B transcription factor. Mol. Cell. Biol.
10:2327-2334.
40. Kunsch, C., R. K. Lang, C. A. Rosen, and M. F. Shannon. 1994. Synergistic
transcriptional activation of the IL-8 gene by NF-kappa B p65 (RelA) and NFIL-6. J. Immunol. 153:153-164.
41. Marui, N., M. K. Offermann, R. Swerlick, C. Kunsck, C. A. Rosen, M.
Ahmad, R. W. Alexander, and R. M. Medford. 1993. Vascular cell adhesion
molecule-i (VCAM-1) gene transcription and expression are regulated through
an antioxidant-sensitive mechanism in human vascular endothelial cells. J. Clin.
Invest. 92:1866-1874.
42. Schreck, R., P. Rieber, and P. A. Baeuerle. 1991. Reactive oxygen intermediates as widely used messengers in the activation of the NF-KB transcription
factor and HIV-1. EMBO (Eur. Mol. Biol. Organ.) J. 10:2247-2258.
43. Huie, R. E., and S. Padmaja. 1993. Reaction of nitric oxide with superoxide
anion. Free Radical Res. Commun. 18:195-200.
44. Schreck, R., K. Albermann, and P. A. Baeuerle. 1992. Nuclear FactorKB: an oxidative stress-responsive transcription factor of eukaryotic cells. Free
Radical Res. Commun. 17:221-237.
45. Kurose, I., P. Kubes, R. Wolf, D. C. Anderson, J. Paulson, M. Miyasaka,
and D. N. Granger. 1993. Inhibition of nitric oxide production. Mechanisms of
vascular albumin leakage. Circ. Res. 73:164-171.
46. Cooke, J. P., and P. S. Tsao. 1994. Is NO an endogenous antiatherogenic
molecule? Arterioscler. Thromb. 14:653-655.